Hi, I found a solution:
>> The setup seems to work and the system vms get deployed on one of the >> xenserver vm, but the SSVM fails to work. I ssh into the SSVM using it's >> link-local IP via its host xenserver's dom0, and find that the SSVM is >> unable to reach the gateway, 192.168.0.1. All routes, iptables look fine, >> comments? > > Which bridge is used by the system vm ? The system vm, specifically ssvm, uses four vifs, one as link-local eth0 with xapi1 interface and other three (eth1,eth2,eth3) with xenbr0. I'm able to ping on ssvm's eth1 from xenserver and vice-versa. And also from my host os to my xenserver and vice versa. When I ping from my ssvm to my gateway (192.168.0.1, host os), it fails. The tcpdump on these interfaces show that arp requests from ssvm reach the gateway and replies are sent and are very slow. And no icmp reaches the gateway from ssvm and vice-versa. The host-only adapter attached to XenServer vm was set by default in promiscuous: deny mode. So, we need to set the host-only adapter's promiscuous mode to 'allow all' and configure global settings (in my case local storage to 'true', mgmt host '192.168.0.1' otherwise ssvm fails) as needed. This should work with multiple devclouds as well. > "brctl show" on each of your hyper visor vm Detailed log of xenserver vm, jfyi: http://pastebin.com/1htDq44J Regards, Rohit
